Beruflich Dokumente
Kultur Dokumente
Website: www.fe.up.pt/vibest
Abstract
This paper presents a brief characterization of the evolution of Experimental Modal
Analysis in the Civil Engineering field, from Input-Output to Output-Only Modal
Identification Techniques, taking particularly into account the experience of the
authors at the Laboratory of Vibrations and Monitoring (VIBEST, www.fe.up.pt/vibest)
of FEUP.
Key words:
1. Introduction
Some decades ago, the major concern of Structural Engineers was the development
and automatic application of new and powerful numerical methods for the analysis
(static and dynamic) and design of large Civil Engineering structures. In this context,
the fast development of the finite element techniques accompanied by the
tremendous technological progress in the field of personal computers allowed the
structural designer to use currently excellent structural analysis software packages,
which enable to accurately simulate the structural behaviour.
However, the design and construction of more and more complex and ambitious civil
structures, like dams, large cable-stayed or suspension bridges, or other special
structures, made structural engineers feel the necessity to develop also appropriate
experimental tools that might enable the accurate identification of the most relevant
structural properties (static and dynamic), providing reliable data to support the
calibration, updating and validation of the structural analysis numerical models used
at the design stage.
Beyond that, the continuous ageing and subsequent structural deterioration of a large
number of existing structures made structural engineers gradually more interested in
the development and application of efficient vibration based damage detection
techniques supported by structural health monitoring systems, in which the regular
identification of modal properties plays also an important role.
Therefore, the first and natural tendency of Civil Engineering researchers was to take
some profit from important previous developments made in System Identification and
Experimental Modal Analysis in Electrical and Mechanical Engineering, trying to
accurately identify the main dynamic properties of civil structures by applying well
established input-output modal identification techniques.
The difficulty to excite large civil structures in a controlled form, as well as remarkable
technological progresses registered in the area of transducers and analogue to digital
www.samco.org
Page 1 of 22
The controlled excitation of large Civil Engineering structures requires however the
use of specific and heavy excitation equipment. One option frequently used in the
www.samco.org
Page 2 of 22
(a)
(b)
(c)
Fig. 2. Servo-hydraulic shakers to excite (a) bridges (vertically), (b) dams (laterally) (EMPA);
(c) Electro-hydraulic shaker from Arsenal Research.
(a)
(b)
(c)
(d)
Fig. 3. Schematic cross-section of (a) piezoelectric, (b) piezoresistive, (c) capacitive and
(d) force balance accelerometers
The electrical signals captured by these transducers are usually rather low and so
must be amplified by conditioning units that may also provide anti-aliasing low-pass
filtering (allowing lower sampling rates) and analogue integration to velocities or
displacements.
The data acquisition and storage of measurement data involves the use of an
analogue-to-digital (A/D) converter inserted in a digital computer. The digital raw data
must be preliminary analysed and processed, considering operations of scale
conversion, trend-removal and decimation. Afterwards, the acceleration time series
can be multiplied by appropriate time windows (Hanning, Cosine-Taper, etc.), in
order to reduce leakage effects, and subdivided in different blocks for evaluation of
average spectral auto and cross spectra estimates, using the FFT algorithm. At last,
estimates of FRFs can be obtained using estimators H1 or H2 [1]. The automatic
www.samco.org
Page 3 of 22
www.samco.org
Page 4 of 22
(a)
(b)
(c)
(a)
(b)
(c)
Fig. 4. (a) Jindo cable-stayed bridge; (b) Physical model on shaking table (EERC, Univ.
Bristol); (c) Physical model on shaking table (ISMES).
Shaker
Exp.
1.00E-01
Ajust.
1.00E-02
8
10
11
12
Frequncia (Hz)
(a)
(b)
Fig. 6. (a) Amplitude of FRF relating vertical acceleration at 1/3 span with the vertical force
applied at the opposite 1/3 span; (b) Identified pattern of a set of multiple modes.
Several large Civil Engineering structures, like buildings, bridges or dams, have been
also submitted to forced vibration tests in the past, using heavy excitation devices
only available at important and well equipped laboratories. That was the case of
EMPA, where Cantieni and other researchers have tested a significant number of
bridges and dams [4-6]. Figures 7-9 show some examples of that remarkable
activity, presenting in particular some of the modes of vibration accurately identified
at the sweedish Norsj dam.
www.samco.org
Page 5 of 22
(a)
(b)
(c)
Fig. 7. (a) Dala bridge; (b) Aarburg bridge; (c) Electro-hydraulic vibrator used at Aarburg
bridge
(a)
(b)
(c)
Fig. 8. (a) Norsj dam; (b,c) View of instrumented point at downstream side reinforced
concrete wall
Fig. 9. Some identified modes of vibration at Norsj dam (modes 1, 2, 3, 10, 11, 12).
www.samco.org
Page 6 of 22
(a)
(b)
Fig. 10. Forced vibration tests of (a) Tatara cable-stayed bridge (http://www.hsba.go.jp)
and (b) Yeongjong suspension bridge (http://www.yeongjongbridge.com)
Fig. 11. (a) Force balance accelerometers; (b) Multi-channel data acquisition and processing
system for ambient vibration tests; (c) Strong motion tri-axial seismograph
www.samco.org
Page 7 of 22
www.samco.org
Page 8 of 22
Welch method
FFT
estimates of
RD functions
RD method
fi
SVD
Dy(t)
y (t)
direct method
LS
QR
orthogonal decomposition
SSI-COV method
Ry(t)
FFT
LS, EVD
LS, EVD
RD-PP method
Sy (f)
response
time series
modal
parameters
SSI-DATA method
QR, SVD, LS, EVD
Very recently, the new operational Polymax parameter estimation method was
introduced by LMS (www.lms.be) [27]. This method operates on spectra or half
spectra (i.e. the Fourier transforms of the positive time lags of the correlation
functions) and its main advantage consists in yielding extremely clear stabilization
diagrams, making an automation of the parameter identification process rather
straightforward, which may enable the continuous monitoring of structural dynamic
properties.
Examples of ambient vibration tests
1.00E+00
Half-sum
ANPSD (N-S)
1.00E-01
Half-diff.
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
0
10
12
10
12
1.00E+00
Mode 1
ANPSD (E-W)
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
Frequency (Hz)
(a)
(b)
Mode 4
(c)
Fig. 13. (a) Heritage Court Tower; (b) ANPSDs spectra; (c) Two identified mode shapes.
www.samco.org
Page 9 of 22
dB | 1.0 / Hz
State Space
Dimension
Stabilization Diagram
Data Set: Deck nodes 4 and 10
PC [Data Driven]
20
83
0
78
-20
73
-40
68
-60
0.3
0.6
0.9
1.2
0.3
Frequency [Hz]
ARTeMIS Extractor, 0f 9-4eaa-9649-64e1, ARTX-0330E-270803PRO, Enterprise License
0.6
0.9
1.2
Frequency [Hz]
ARTeMIS Extractor, 0f 9-4eaa-9649-64e1, ARTX-0330E-270803PRO, Enterprise License
Fig. 16. Identified first (left) and second (right) bending (up) and torsion (down) modes.
Ambient vibration tests have been performed with great success in large buildings,
bridges or other special structures, allowing the creation of high quality experimental
databases that have been used to compare the performance of different output-only
modal identification methods. In this context, a benchmark test, concerning the modal
identification of the Heritage Court Tower (Vancouver, Canada) has been organized
at IMAC-XVIII by Ventura [28]. This example could show the interest of considering
a pre-combination of measured signals (half-sum and half-difference signals along
two orthogonal directions at two different points at each floor) to emphasize the
contributes from bending or torsion, as well as of a high frequency resolution to
separate contributions from close modes, when using the classical PP method [29].
The application of FDD and SSI methods permitted a more automatic identification
procedure, distinguishing close modes and extracting modal damping estimates.
In the field of bridges, complete ambient vibration tests were developed along about
5km of the Vasco da Gama Bridge by FEUP. Regarding the main cable-stayed
www.samco.org
Page 10 of 22
1E-03
1E-04
1E-05
1E-06
0.25
0.35
0.45
Frequency (Hz)
(a)
Power Spectral Density functions at reference section
(half-sum of vertical acceleration)
Amplitude
1E-02
0.55
0.65
(b)
Power Sepctral Density functions at reference section
(half-sum of transversal acceleration)
Amplitude
1E-02
1E-03
1E-03
1E-04
1E-04
1E-05
1E-05
1E-06
wind - 2 m/s
1E-06
0
0.2
0.4
0.6
0.8
1.2
wind - 10 m/s
(c)
2.2
wind - 14 m/s
2.4
2.6
2.8
wind - 2 m/s
1E-07
3
0.3
0.6
0.9
wind - 10 m/s
1.2
1.5
1.8
Frequency (Hz)
2.1
wind - 14 m/s
2.4
2.7
(d)
Fig. 17. (a) Guadiana cable-stayed bridge; PSD functions of half-sum vertical (b, c)
and lateral (d) acceleration at reference section, as function of mean wind speed.
Its worth noting that the existence of cables frequencies in the frequency range of
analysis can difficult the identification of global natural frequencies. Figures 17 (c,d)
show, for instance, PSD functions concerning the ambient response of the
international Guadiana cable stayed-bridge (Fig. 17 (a)) (linking Portugal to Spain in
Algarve), evaluated with three different levels of average wind speed, which show the
appearance of spectral contributes from the fundamental modes of stay-cables (in
the range 0.6-0.9Hz) or second harmonics, leading to spectral peaks that can not be
interpreted as global natural frequencies of the bridge. Inspection of the spectral
peaks (Figure 17 (b)) shows also the increase of modal damping with wind speed,
which can be evaluated through ambient vibration tests using sufficiently long
measurement periods [33].
Its still worth mentioning that the above described output-only modal identification
technique used in Vasco da Gama and Guadiana bridges by FEUP, has been also
recently applied with great success in the dynamic tests at the commissioning stage
of the outstanding Millau viaduct, coordinated by Grillaud and Flamand (CSTB,
France) [34].
The same approach can be used in the experimental assessment of other special
structures, like innovative stadia roofs, particularly susceptible to wind induced
vibrations, as consequence of their slenderness. That is, for instance, the case of the
new Braga Sports Stadium suspended roof (EURO2004) (Figure 18), formed by 68
www.samco.org
Page 11 of 22
www.samco.org
Page 12 of 22
f = 0.293 Hz
f = 0.523 Hz
f = 0.540 Hz
f = 0.559 Hz
f = 0.632 Hz
f = 0.654 Hz
f = 0.701 Hz
f = 0.874 Hz
www.samco.org
Page 13 of 22
(b)
(a)
(c)
Fig. 20. (a) Free vibration test of Vasco da Gama cable-stayed bridge; (b) Aerial view of
Madeira airport extension; (c) Mass of 60.8t used in the free vibration test of Madeira
airport
In the cable-roof of the new Braga Stadium (EURO2004) (Fig. 18), whose
aeroeleastic stability was proved by different experimental tests on physical models,
the modal damping identification was essentially required to study possible
resonance effects, which may affect the structural integrity and durability in the long
term.
The identification of modal damping ratios was developed, in a first instance [37]
based on a set of data collected by the instrumentation installed at the roof structure,
during the forced and free vibration tests developed at commissioning phase.
The free vibration test was based on the sudden release of a 5 ton mass from the
roof, the structural response (Figure 21a) being collected by the six tri-axial forcebalance accelerometers of the dynamic monitoring system. The application of bandpass filters to the measured signals enables the evaluation of modal free decay
responses; these were used to estimate the modal damping coefficients presented in
the second column of Table 2. This procedure faces two problems, which justify the
missing values in the table: the low level of excitation of some modes and the
difficulty of isolate the contribution of modes with very close natural frequencies. An
alternative to this procedure is to use the measured response to the impulse as input
to the SSI-COV method, as the responses to impulses are proportional to the
correlations of the responses to a white noise excitation. This technique was used
after the application of a low-pass filter, with a cut-off frequency of 1 Hz, and a
decimation to reduce the sampling frequency to 5 Hz, and provided the results
presented in the third column of Table 2. It is interesting to observe that for the
modes where both techniques were applied the results are very consistent.
Forced vibration tests were further conducted, based on a harmonic excitation of the
roof at resonance, by means of a cable pulled by an electric engine. After resonance
was attained, the excitation was suppressed and the free vibration response
measured at the same 6 accelerometers. Figure 21b) represents one of the
measured free decays. Using this procedure 5 modes were excited and so, 5 free
www.samco.org
Page 14 of 22
(b)
(a)
Fig. 21. Free decay (a) after application of an impulse and (b) after excitation of the 2nd mode
The comparison between the modal damping coefficients identified using artificial
and ambient excitation shows the existence of a satisfactory correlation. However,
one can notice that relative differences tend to increase at lower frequencies. In
effect, it is well known that it is very difficult to estimate modal damping coefficients
since they are dependent on the amplitude of vibration and also on the wind
characteristics because of the existence of aeroelastic damping. The very low
damping values of this structure make the comparison even more difficult, because
very small differences are expressed by significant relative errors.
It is important to stress that, in this very flexible structure, the results provided by the
FDD method are comparable with the ones of the SSI methods due to the very long
time series used and because it was adopted an alternative procedure [35] to
estimate the correlation function. The application of the standard EFDD method using
independently the time series of each setup (with 16 min.) led to values of modal
damping coefficients for the first modes of about 1%.
Table 2. Summary of all the identified modal damping coefficients (%).
Mode
1
2
3
4
5
6
7
8
9
10
11
Free Vib.
Filter
0.28
0.25
0.34
0.20
-
Free Vib.
SSI-COV
0.29
0.37
0.32
0.22
0.44
0.36
0.29
0.11
0.18
0.18
-
Harmonic
excitation
0.28
0.27
0.22
0.43
0.20
-
Amb.
FDD
0.58
0.52
0.47
0.35
0.25
0.36
vibration
SSI-COV
0.50
0.42
0.44
0.40
0.47
0.54
0.28
0.27
0.26
0.26
0.41
SSI-DATA
0.51
0.48
0.39
0.33
0.53
0.47
0.73
0.30
0.31
0.29
0.38
www.samco.org
Page 15 of 22
Fig 22. Lateral view and schematic of the New Hintze Ribeiro Bridge, over Douro river, with
indication of the measurement points used along the deck.
In the first case (Fig. 22), rather good correlation between identified and calculated
modal parameters was achieved concerning the vertical bending modes of vibration
(Table 3, left). However, regarding the lateral response of the bridge, identified
frequencies were systematically considerably higher than the corresponding
calculated values (Table 3, right), though good correlation of modal shapes has been
obtained [39]. Such discrepancy stems naturally from the difficulty to simulate
numerically the real characteristics of soil-structure interaction at the foundation of
the several piers, and it shows the large influence that variations of boundary
conditions can have on the global dynamic bridge properties.
www.samco.org
Page 16 of 22
Identified
frequency
(Hz)
1.465
1.782
Type of
mode
2.291
2.711
2.197
2.710
3rd vertical
4th vertical
3.458
3.54
5th vertical
1st vertical
2nd vertical
Calculated
frequency
(Hz)
0.715
0.892
1.180
Identified
frequency
(Hz)
1.147
1.636
2.881
Type of
mode
1st lateral
2nd lateral
3rd lateral
In the case of Pinho Bridge (Fig. 23), very similar modal estimates were obtained in
the three similar spans and good correlation was achieved between significant
identified and calculated modal parameters, considering either the vertical or the
lateral behaviour of the bridge. However, it was clearly noticed [40] that the initial
numerical modeling developed by the designer should be improved to correctly
simulate the lateral dynamic response by including the stiffness associated to the
concrete slab of the deck, which was made through a discretization in shell elements.
Table 4. Comparison of calculated and identified natural frequencies.
Mode
Vertical
bending
Lateral
bending
Torsion
type
1st
2nd
3rd
4th
5th
1st
2nd
3rd
1st
Experimental
2.779
5.460
8.293
10.604
12.577
1.721
3.210
4.273
5.937
Frequency (Hz)
Model_1
2.380
4.676
6.352
7.383
8.106
1.165
2.020
4.278
5.087
Model_2
2.998
5.447
8.695
11.749
13.396
1.758
3.523
4.495
6.921
This fact is evidenced by Table 4, where the identified natural frequencies are
compared with the calculated ones using either a numerical model exclusively based
on a discretization in bar elements (Model_1) or an improved model including the
discretization of the concrete slab in shell elements (Model_2).
www.samco.org
Page 17 of 22
For the purpose of finite element updating, a complete ambient vibration test was
performed and initial finite element models were developed idealizing the bridge deck
as a set of beam elements with the geometry considered at the design stage or
measured through a topographic survey (Models 1 and 2). Afterwards, due to the
clear nonlinear geometrical behaviour of the bridge, a third model (Model 3) was
developed discretizing the deck in truss finite elements with the cables axial stiffness
(neglecting bending stiffness), adjusting the initial cables tension so as to obtain the
measured longitudinal profile after progressive application of the loads. In order to
take also into account the bending stiffness of the concrete slab, this model was
subsequently adapted (Model 4), discretizing the deck in truss finite elements, with
progressive loading and activation of beam elements connecting the nodes of the
truss elements, simulating the effective construction procedure.
Finally, this model was slightly modified, considering partial rotations between beam
elements to simulate the lack of sealing of the joints and reducing the area and
inertia of the beam elements to simulate the effects of cracking and lack of
adherence between precast and cast in situ concrete. After all these iterations, very
good level of correlation between identified and calculated natural frequencies and
mode shapes was achieved, as extensively described in ref. [41].
Subsequently, similar finite element updating procedure was also performed
considering the three-dimensional nature of the bridge and discretizing the concrete
slab in shell elements (Figure 25), which led also to a good correlation of calculated
and identified natural frequencies (Table 5) and mode shapes [42].
www.samco.org
Page 18 of 22
Calculated
frequency
(Hz)
0.940
2.101
2.152
2.234
2.344
3.736
3.875
4.188
4.200
5.761
Identified
frequency
(Hz)
0.972
2.072
2.043
1.960
2.382
3.590
4.165
5.461
Identified
damping
(%)
1.07
1.51
1.40
1.10
1.72
1.78
2.00
1.92
Type of mode
First symmetric (two spans, opposite phase)
Second anti-symmetric (L=30m)
First symmetric (two spans, in-phase)
First lateral (two spans)
Second anti-symmetric (L=28m)
Second symmetric (L=30m)
First torsional (two spans, L=30m dominant)
Second symmetric (L=28m)
First torsional (two spans, L=28m dominant)
Third anti-symmetric (L=30m)
F=2.101Hz
F=2.152Hz
F=2.234Hz
F=2.344Hz
Beyond this type of sensitivity analyses, more automatic finite element updating
techniques can also be used [43]. In this context, a drawback of output-only modal
identification seemed to be the impossibility to obtain mass normalized mode shapes.
However, this inconvenient can be overcome [44] introducing appropriate mass
changes.
5. Conclusion
Civil Engineering structures have peculiar characteristics (large size and relatively
low natural frequencies) that make difficult the current application of classical inputoutput modal identification techniques. Therefore, there is presently a clear tendency
worldwide to explore and improve the potential of output-only modal identification
techniques, whose efficiency and accuracy were clearly illustrated with the
applications shown. These techniques, that may be used under normal operation
conditions, can provide a solid basis for: (i) the development of finite element
correlation analyses, (ii) the finite element updating and validation; (iii) the definition
www.samco.org
Page 19 of 22
6. References
[1]
[2]
Han, M-C & Wicks, A.L. On the application of Forsythe orthogonal polynomials
th
for global modal parameter estimation, Proc. 7 Int. Modal Analysis Conference,
1989
[3]
[4]
[5]
Cantieni, R., Deger, Y. & Pietrzko, S. Large structure investigation with dynamic
methods: the bridge on the river Aare at Aarburg, Prestressed Concrete in
th
Switzerland, Report of the Swiss FIP Group to the 12 FIP Congress,
Washington D.C., 1994
[6]
[7]
[8]
[9]
[10]
[11]
Mc Lamore, V.R., Hart, G. & Stubbs, I.R. Ambient Vibration of Two Suspension
Bridges, Journal of the Structural Division, ASCE, Vol.97, N.ST10, p.2567-2582,
1971
[12]
[13]
[14]
[15]
Corra, M.R. & Campos Costa, A. Ensaios Dinmicos da Ponte sobre o Rio
Arade, in Pontes Atirantadas do Guadiana e do Arade (in Portuguese), ed. by
LNEC, 1992
www.samco.org
Page 20 of 22
[17]
[18]
Brincker, R., Krenk, S., Kirkegaard, P.H. & Rytter, A. Identification of the
Dynamical Properties from Correlation Function Estimates, Bygningsstatiske
Meddelelser, Danish Society for Structural Science and Engineering, Vol.63,
N.1, p.1-38, 1992
[19]
[20]
Ibrahim, S.R. & Mikulcik, E.C. A Method for the Direct Identification of Vibration
Parameters from the Free Response, The Shock and Vibration Bulletin, Vol.47,
N.4, p.183-198, 1977
[21]
[22]
[23]
Vold, H., Kundrat, J., Rocklin, G.T. & Russel, R. A Multi-Input Modal Estimation
Algorithm for Mini-Computers, SAE Technical Paper Series, N.820194, 1982
[24]
[25]
[26]
[27]
Peeters, B., Vanhollebeke, F. & Van der Auweraer, H. Operational PolyMAX for
Estimating the Dynamic Properties of a Stadium Structure during a Football
rd
Game, Proc. 23 Int. Modal Analysis Conference, Orlando, USA, 2005
[28]
[29]
[30]
Peeters, B., De Roeck, G., Caetano, E. & Cunha, A. Dynamic study of the
Vasco da Gama Bridge, Proc. of the International Conference on Noise and
Vibration Engineering, ISMA, Leuven, Belgium, 2002
[31]
Cunha, A., Caetano, E., Brincker, R. & Andersen, P. Identification from the
nd
Natural Response of Vasco da Gama Bridge, Proc. 22 Int. Modal Analysis
Conference, Deaborn, USA, 2004
[32]
[33]
www.samco.org
Page 21 of 22
[35]
[36]
Rodrigues, J. & Campos Costa, A. Dynamic Tests of the Structure for Extension
th
of the Madeira Island Airport, Proc. 20 Int. Modal Analysis Conference, L.A.,
USA, 2002
[37]
[38]
[39]
Caetano, E. & Cunha, A. Ambient Vibration Test and Finite Element Correlation
of the New Hintze Ribeiro Bridge, Proc. Int. Modal Analysis Conf., Kissimmee,
USA, 2003
[40]
[41]
[42]
[43]
[44]
www.samco.org
Page 22 of 22